Volume 51, Number 4

Drug-Induced Rheumatic Syndromes

Raymond Yung, MD
Department of Internal Medicine
University of Michigan
Ann Arbor, MI

Bruce Richardson, MD, PhD
Department of Internal Medicine
Ann Arbor Veterans Affairs Medical Center
Ann Arbor, MI

Drug-induced Lupus (DIL)
The incidence of DIL has been estimated to be ~15,000 to 20,000 per year with ~30,000 to 50,000 patients being affected in the United States at any given time (2). More than 100 drugs have now been implicated in DIL (Table 1).

The drugs that have been most studied in DIL are procainamide and hydralazine (2). However, these drugs are not commonly prescribed today and the lupus-like illness they produce is often different from those implicated in DIL in recent years. Unlike patients with idiopathic lupus, patients with hydralazine-induced lupus (HIL) and procainamide-induced lupus (PIL) are usually older and more likely to be men, reflecting the age and sex of the patients receiving these drugs.  One-third of patients receiving procainamide for more than a year will develop symptoms, and almost all will become ANA-positive after 2 years. In contrast, less than 20% of hydralazine-treated patient will develop DIL. The patients at the greatest risk for developing HIL are those taking doses greater than 200 mg/day and those who have received more than 100 grams cumulative dose.

Compared to idiopathic lupus patients, DIL patients have less renal, neuropsychiatric, and skin manifestations. Pleuropulmonary complaints are particularly common in PIL. More than 90% of PIL and HIL patients have antihistone antibodies, and 20% to 40% have rheumatoid factors. By definition, patients with DIL are ANA-positive. The pattern of the ANA is usually homogenous, reflecting reactivity again histone or histone-DNA complexes.  Accordingly, a speckled pattern ANA in a patient with suspected HIL or PIL makes a drug-induced etiology less likely.

A number of drugs have been reported to induce antiphospholipid antibodies (Table 2). The strongest associations are with the use of chlorpromazine and procainamide. Although there are occasional reports of thrombosis, the risk of an associated clinical event appears to be low. For example, 56% to 87% of patients on chronic chlorpromazine therapy develop antiphospholipid antibodies but less than 6% of the antibody-positive patients have a thrombotic episode in up to 5 years follow-up (3,4). The reason for the low incidence of thrombosis is unclear, but most of the drug-induced anticardiolipin antibodies belong to the IgM subtype and probably have lower pathogenic potential.

Minocycline is a popular acne treatment and is also used in the treatment of rheumatoid arthritis.  Over the past decade, numerous reports of immune and autoimmune phenomena associated with its use have appeared in the literature, including serum sickness, lupus, autoimmune hepatitis, and vasculitis (5,6).  Women are more likely to be affected.  Symptoms also tend to be milder than idiopathic lupus, and include fever, arthralgia/myalgias and arthritis.  Renal and central nervous system involvement again are rare. Pneumonitis (included in the first reported case of minocycline-induced lupus) and hepatitis (in isolation or as part of the lupus syndrome) have been described in a number of cases, although the true incidence is unknown.  In addition, there are rare reports of skin disorders, including livido reticularis and cutaneous vasculitis. Serological changes in minocycline-induced lupus are also different from classical DIL with a higher incidence of anti-dsDNA antibodies (up to 30%), a positive p-ANCA (70%), and a lower incidence of antihistone (<20%) antibodies.

A number of the new recombinant biologics have been implicated in DIL including interferon a (IFNa), interferon g (IFNg), and anti-TNF therapies (7).  Between 4% and 19% of IFNa-treated patients develop some forms of autoimmunity (7,8,9,10).  Approximately 12% of patients receiving this drug develop an ANA, and between 0.15% and 0.7% will develop a lupus-like illness.  Anti-dsDNA antibodies occur in 8%, and almost all with IFNa-induced lupus have elevated anti-dsDNA antibodies. Although female gender appears to be a risk factor for the development of anti-dsDNA antibodies, approximately equal numbers of cases of male and female IFNa-induced lupus are reported in the literature.  The duration of treatment to onset of lupus ranges from 1 month to 7 years; the dose and duration may be important determinants. Because autoimmune thyroid diseases are common following IFNa treatment (especially in those with pre-existing thyroid antibodies), patients may also present to the clinician with soft-tissue rheumatic complaints secondary to thyroid dysfunction.

Anti-TNF therapy represents a significant advance in the treatment of rheumatoid arthritis. It has also become apparent that patients receiving either anti-TNF monoclonal antibodies or TNF receptor antagonists can develop serological evidence of autoimmunity (11).  In one study, new ANA and anti-dsDNA antibodies were found in 33% and 9% of infliximab recipients, respectively.  However, there is clear discordance between the development of autoantibodies and clinical autoimmunity, with less than 20 symptomatic patients reported in the literature and in abstract form.

At present, it is not possible to predict who will develop DIL.  If necessary, patients with idiopathic lupus should be allowed to take potentially lupus-inducing drugs, including procainamide and hydralazine, but with careful monitoring.